Crystal Structure of Mammalian Cysteine Dioxygenase

Cysteine dioxygenase is a mononuclear iron-dependent enzyme responsible for the oxidation of cysteine with molecular oxygen to form cysteine sulfinate. This reaction commits cysteine to either catabolism to sulfate and pyruvate or the taurine biosynthetic pathway. Cysteine dioxygenase is a member of the cupin superfamily of proteins. The crystal structure of recombinant rat cysteine dioxygenase has been determined to 1.5-Å resolution, and these results confirm the canonical cupin β-sandwich fold and the rare cysteinyltyrosine intramolecular cross-link (between Cys93 and Tyr157) seen in the recently reported murine cysteine dioxygenase structure. In contrast to the catalytically inactive mononuclear Ni(II) metallocenter present in the murine structure, crystallization of a catalytically competent preparation of rat cysteine dioxygenase revealed a novel tetrahedrally coordinated mononuclear iron center involving three histidines (His86, His88, and His140) and a water molecule. Attempts to acquire a structure with bound ligand using either cocrystallization or soaking crystals with cysteine revealed the formation of a mixed disulfide involving Cys164 near the active site, which may explain previously observed substrate inhibition. This work provides a framework for understanding the molecular mechanisms involved in thiol dioxygenation and sets the stage for exploration of the chemistry of both the novel mononuclear iron center and the catalytic role of the cysteinyl-tyrosine linkage. Cysteine dioxygenase is a mononuclear iron-dependent enzyme responsible for the oxidation of cysteine with molecular oxygen to form cysteine sulfinate. This reaction commits cysteine to either catabolism to sulfate and pyruvate or the taurine biosynthetic pathway. Cysteine dioxygenase is a member of the cupin superfamily of proteins. The crystal structure of recombinant rat cysteine dioxygenase has been determined to 1.5-Å resolution, and these results confirm the canonical cupin β-sandwich fold and the rare cysteinyltyrosine intramolecular cross-link (between Cys93 and Tyr157) seen in the recently reported murine cysteine dioxygenase structure. In contrast to the catalytically inactive mononuclear Ni(II) metallocenter present in the murine structure, crystallization of a catalytically competent preparation of rat cysteine dioxygenase revealed a novel tetrahedrally coordinated mononuclear iron center involving three histidines (His86, His88, and His140) and a water molecule. Attempts to acquire a structure with bound ligand using either cocrystallization or soaking crystals with cysteine revealed the formation of a mixed disulfide involving Cys164 near the active site, which may explain previously observed substrate inhibition. This work provides a framework for understanding the molecular mechanisms involved in thiol dioxygenation and sets the stage for exploration of the chemistry of both the novel mononuclear iron center and the catalytic role of the cysteinyl-tyrosine linkage. The cytosolic enzyme cysteine dioxygenase (CDO) 3The abbreviations used are: CDO, cysteine dioxygenase; SAD, single-wavelength anomalous diffraction; PDB, Protein Data Bank; QDO, quercetin dioxygenase. 3The abbreviations used are: CDO, cysteine dioxygenase; SAD, single-wavelength anomalous diffraction; PDB, Protein Data Bank; QDO, quercetin dioxygenase. (EC 1.13.11.20) catalyzes the irreversible oxidation of cysteine to cysteine sulfinate (Reaction 1). This reaction is required for a variety of critical metabolic pathways (1Stipanuk M.H. Annu. Rev. Nutr. 2004; 24: 539-577Crossref PubMed Scopus (734) Google Scholar). CDO initiates the catabolism of cysteine to pyruvate and sulfate, which is essential for the provision of adequate inorganic sulfate and allows pyruvate to enter central pathways of metabolism. Also, the oxidation and excretion of the sulfur of methionine depends on CDO, because the sulfur atoms of methionine and homocysteine are only oxidized after their transfer, via the transsulfuration pathway, to serine to yield cysteine. In addition, CDO activity is essential for the biosynthesis of taurine, which is formed by the decarboxylation of cysteine sulfinate to hypotaurine and further oxidation of hypotaurine to taurine. Clinical evidence indicates that a block in cysteine catabolism, thought to be at CDO, leads to an altered cysteine to sulfate ratio that is associated with sulfate depletion and other adverse effects (1Stipanuk M.H. Annu. Rev. Nutr. 2004; 24: 539-577Crossref PubMed Scopus (734) Google Scholar). The prevalence of impaired cysteine catabolism has been reported to be increased in patient populations afflicted with rheumatoid arthritis, liver diseases, Parkinson disease, Alzheimer disease, motor neuron disease, and systemic lupus erythematosus (2Heafield M.T. Fearn S. Steventon G.B. Waring R.H. Williams A.C. Sturman S.G. Neurosci. Lett. 1990; 110: 216-220Crossref PubMed Scopus (374) Google Scholar, 3Heafield M.T. Williams A.C. Curr. Opin. Neurol. Neurosurg. 1992; 5: 288-294PubMed Google Scholar, 4Bradley H. Gough A. Sokhi R.S. Hassell A. Waring R. Emery P. J. Rheumatol. 1994; 21: 1192-1196PubMed Google Scholar, 5Davies M.H. Ngong J.M. Pean A. Vickers C.R. Waring R.H. Elias E. J. Hepatol. 1995; 22: 551-560Abstract Full Text PDF PubMed Scopus (36) Google Scholar). These patients frequently exhibit low levels of sulfate in plasma (and in synovial fluid), elevated fasting plasma cysteine concentrations, elevated plasma cysteine to sulfate ratios, and an impaired capacity for sulfation reactions in vivo. Reduced cysteine catabolism would cause both depletion of the products sulfate and taurine and an accumulation of the substrate cysteine, either of which would lead to adverse effects. Large doses of cysteine or cystine have been shown to be toxic in several species (6Pedersen O.O. Karlsen R.L. Investig. Ophthalmol. Vis. Sci. 1980; 19: 886-892PubMed Google Scholar, 7Karlsen R.L. Grofova I. Malthe-Sorenssen D. Fonnum F. Brain Res. 1981; 208: 167-180Crossref PubMed Scopus (78) Google Scholar, 8Mathisen G.A. Fonnum F. Paulsen R.E. Neurochem. Res. 1996; 21: 293-298Crossref PubMed Scopus (26) Google Scholar). Cysteine is thought to be neuroexcitotoxic, acting via effects on glutamate transport by systems mathrmXC− and mathrmXAG− and on the N-methyl-d-aspartate subtype of the glutamate receptor (9McBean G.J. Flynn J. Biochem. Soc. Trans. 2001; 29: 717-722Crossref PubMed Google Scholar, 10Schubert D. Piasecki D. J. Neurosci. 2001; 21: 7455-7462Crossref PubMed Google Scholar), and cysteine can form toxins by reacting with other compounds (11Li H. Dryhurst G. J. Neurochem. 1997; 69: 1530-1541Crossref PubMed Scopus (116) Google Scholar). Taurine status is associated with sulfur amino acid intake and thus with its synthesis from cysteine, and a lack of adequate taurine has been associated with a number of abnormalities, most commonly with dilated cardiomyopathy, impaired neurological development, and retinal photoreceptor cell abnormalities and photoreceptor cell death (12Huxtable R.J. Physiol. Rev. 1992; 72: 101-163Crossref PubMed Scopus (2240) Google Scholar). CDO was first described by Ewetz and Sorbo (13Ewetz L. Sorbo B. Biochim. Biophys. Acta. 1966; 128: 296-305Crossref PubMed Scopus (57) Google Scholar), who postulated that it might be a mixed function oxidase. Subsequently, Lombardini et al. (14Lombardini J.B. Singer T.P. Boyer P.D. J. Biol. Chem. 1969; 244: 1172-1175Abstract Full Text PDF PubMed Google Scholar) demonstrated that the enzyme was a dioxygenase and did not require NAD(P)H as an electron donor. CDO was purified from rat liver by Yamaguchi et al. (15Yamaguchi K. Hosokawa Y. Kohashi N. Kori Y. Sakakibara S. Ueda I. J. Biochem. (Tokyo). 1978; 83: 479-491Crossref PubMed Scopus (58) Google Scholar), who showed it to have a high specificity for cysteine as compared with various cysteine analogs. Little additional work had been done to further characterize the structure or catalytic mechanism of CDO until our recent purification of catalytically active recombinant CDO with kinetic properties that match those observed for CDO in rat liver homogenates: a Km for cysteine of 0.45 mm, a requirement for ferrous ions, and a pH optimum of 6.1 (16Simmons C.R. Hirschberger L.L. Machi M.S. Stipanuk M.H. Protein Expression Purif. 2005; 47: 74-81Crossref PubMed Scopus (48) Google Scholar). This study also demonstrated that recombinant CDO is expressed as both active and inactive isoforms, indicating that significant attention to isolation of the active species would be necessary for structural studies. The function of CDO has been studied most thoroughly in mammals, where it is expressed primarily in liver hepatocytes (17Hirschberger L.L. Daval S. Stover P.J. Stipanuk M.H. Gene (Amst.). 2001; 277: 153-161Crossref PubMed Scopus (27) Google Scholar, 18Stipanuk M.H. Londono M. Hirschberger L.L. Hickey C. Thiel D.J. Wang L. Amino Acids. 2004; 26: 99-106Crossref PubMed Scopus (23) Google Scholar, 19Stipanuk M.H. Hirschberger L.L. Londono M.P. Cresenzi C.L. Yu A.F. Am. J. Physiol. 2004; 286: E439-E448Crossref PubMed Scopus (41) Google Scholar, 20Dominy J.E. Hirschberger Jr., L.L. Coloso R. M. Stipanuk M.H. Biochem. J. 2006; 394: 267-273Crossref PubMed Scopus (66) Google Scholar, 21Stipanuk M.H. Londono M. Lee J.I. Hu M. Yu A.F. J. Nutr. 2002; 132: 3369-3378Crossref PubMed Scopus (109) Google Scholar, 22Hwang L. Hocking-Murray D. Bahrami A.K. Andersson M. Rine J. Sil A. Mol. Biol. Cell. 2003; 14: 2314-2326Crossref PubMed Google Scholar, 23Sacco M. Maresca B. Kumar B.V. Kobayashi G.S. Medoff G. J. Bacteriol. 1981; 146: 117-120Crossref PubMed Google Scholar). In the rat and mouse, CDO is expressed in a highly tissue-specific manner, but CDO abundance in tissues where it is expressed is regulated largely, if not entirely, by cysteine-mediated regulation of CDO degradation (20Dominy J.E. Hirschberger Jr., L.L. Coloso R. M. Stipanuk M.H. Biochem. J. 2006; 394: 267-273Crossref PubMed Scopus (66) Google Scholar, 21Stipanuk M.H. Londono M. Lee J.I. Hu M. Yu A.F. J. Nutr. 2002; 132: 3369-3378Crossref PubMed Scopus (109) Google Scholar). The reaction catalyzed by CDO is notably different from those catalyzed by other classes of dioxygenases that have been studied. First, cysteine dioxygenation involves the oxidation of a sulfhydryl group rather than cleavage of a C-C bond or hydroxylation of a carbon atom, and second, both oxygen atoms from the oxygen molecule are transferred to a single sulfur atom rather than distributed between two carbon atoms. These other classes of dioxygenases are: (i) the Fe(II)-containing vicinal oxygen chelate or type I extradiol dioxygenases that catalyze aromatic ring cleavage of catechols at a C-C bond adjacent to an ortho-hydroxyl substituent; (ii) the Fe(II)/Fe-S-center-containing Rieske dioxygenases that catalyze the cis-hydroxylation of an arene double bond; (iii) the Fe(III)-containing intradiol dioxygenases that cleave aromatic rings between two carbons that each bear a hydroxyl group; (iv) several transition metal-dependent dioxygenases that belong to the cupin superfamily and cleave C-C bonds; and (v) the α-ketoglutarate-dependent Fe(II) dioxygenases (hydroxylases), many of which have also been described as cupins, that couple the oxidative decomposition of α-ketoglutarate to the hydroxylation of a cosubstrate. Mammalian CDO was assigned to the cupin superfamily (24Dunwell J.M. Khuri Sawsan Gane P.J. Microbiol. Mol. Biol. Rev. 2000; 64: 153-179Crossref PubMed Scopus (278) Google Scholar) by the presence of two short but partially conserved sequence motifs, GX5HXHX3-6EX6G and GX5-7PXGX2HX3N, that are separated by 29 residues. Proteins in the cupin family have a wide range of enzymatic and biological functions and often show very low overall sequence similarity but share a canonical cupin “jelly roll” β-barrel (25Dunwell J.M. Culham A. Carter C.E. Sosa-Aguirre C.R. Goodenough P.W. Trends Biochem. Sci. 2001; 26: 740-746Abstract Full Text Full Text PDF PubMed Scopus (253) Google Scholar). Determined as part of a structural genomics effort, the recently reported structure of recombinant Mus musculus CDO-1 (26McCoy J.G. Bailey L.J. Bitto E. Bingman C.A. Aceti D.J. Fox B.G. Phillips Jr. G.N. Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 3084-3089Crossref PubMed Scopus (155) Google Scholar) confirmed the cupin fold, revealed the geometry of the active site when it contains a catalytically incompetent nickel ion, and revealed the presence of a rare cysteinyltyrosine cross-link. To give mechanistic work on CDO a firm foundation, we independently initiated crystallographic studies of recombinant Rattus norvegicus CDO (identical in sequence to mouse CDO). Here, we describe structures at 1.5-Å resolution of both the native iron-containing CDO and a substrate-inhibited complex. The observed iron metallocenter geometry is distinct from that of the nickel center reported for the mouse CDO structure, and this has major ramifications for mechanistic proposals. The active site geometry reported here provides a framework for understanding the molecular mechanisms involved in thiol dioxygenation and sets the stage for exploring the chemistry of this new type of mononuclear iron center. Expression, Purification, and Crystallization of CDO—Native R. norvegicus cysteine dioxygenase (SwissProt/TrEMBL P21816) was prepared as described previously (27Simmons C.R. Hao Q. Stipanuk M.H. Acta Crystallogr. Sect. F. 2005; 61: 1013-1016Crossref PubMed Scopus (9) Google Scholar). The purified protein used for crystallization had a kcat of ∼43 min-1 and a Km of 0.45 mm for l-cysteine when assayed in the presence of ferrous ions (16Simmons C.R. Hirschberger L.L. Machi M.S. Stipanuk M.H. Protein Expression Purif. 2005; 47: 74-81Crossref PubMed Scopus (48) Google Scholar). Expression and purification of selenomethionine (Se-Met)-substituted CDO followed a protocol adapted from Doublié (28Doublié S. Methods Enzymol. 1997; 276: 523-530Crossref PubMed Scopus (794) Google Scholar). BL21(DE3) cells (Novagen) transformed with the pET32a expression vector as described previously (16Simmons C.R. Hirschberger L.L. Machi M.S. Stipanuk M.H. Protein Expression Purif. 2005; 47: 74-81Crossref PubMed Scopus (48) Google Scholar) were grown in M9 salts supplemented with 2 mm MgSO4, 0.4% glucose, 0.002% thiamine (vitamin B1), 0.1 mm CaCl2, and 100 μg/ml carbenicillin at 37 °C to A600 ∼ 0.6. Inhibition of bacterial methionine biosynthesis was targeted by the addition of lysine, phenylalanine, and threonine at 100 mg/liter each, isoleucine, leucine, and valine at 50 mg/liter, and Se-Met at 60 mg/liter. After 15 min, the expression of CDO was induced with 1 mm isopropyl β-d-thiogalactoside, and the cells were incubated at 25 °C overnight. The cells were harvested via centrifugation at 6000 × g for 10 min. Cell lysis and protein purification were performed as described for native CDO (16Simmons C.R. Hirschberger L.L. Machi M.S. Stipanuk M.H. Protein Expression Purif. 2005; 47: 74-81Crossref PubMed Scopus (48) Google Scholar, 27Simmons C.R. Hao Q. Stipanuk M.H. Acta Crystallogr. Sect. F. 2005; 61: 1013-1016Crossref PubMed Scopus (9) Google Scholar), except all buffers were supplemented with 5 mm dithiothreitol to prevent oxidation of the Se-Met. The final protein concentrations of the purified native and Se-Met CDO that were used for crystallization were 7.5 and 6 mg/ml, respectively. Crystallization of native and Se-Met CDO was performed as described (27Simmons C.R. Hao Q. Stipanuk M.H. Acta Crystallogr. Sect. F. 2005; 61: 1013-1016Crossref PubMed Scopus (9) Google Scholar) in sitting drops at 25 °C using a reservoir of 0.1-0.25 m ammonium acetate, 0.1 m tri-sodium citrate, pH 5.6, with 22-26% (w/v) polyethylene glycol 4000 (28Doublié S. Methods Enzymol. 1997; 276: 523-530Crossref PubMed Scopus (794) Google Scholar). Equivalent crystals could also be grown using a reservoir of 0.15 m ammonium sulfate, 0.2 m sodium cacodylate, pH 6.5, with 26% (w/v) polyethylene glycol 8000, and the co-crystal with 5 mm cysteine were grown under these conditions. In all crystallization setups 1.5 μl of concentrated protein solution was mixed with an equal volume of reservoir solution. Cryomounting of CDO crystals was done as described previously (27Simmons C.R. Hao Q. Stipanuk M.H. Acta Crystallogr. Sect. F. 2005; 61: 1013-1016Crossref PubMed Scopus (9) Google Scholar). Crystallographic Data Collection—Crystallographic data collection of native CDO crystals was performed at the National Synchrotron Light Source X12b beamline on an ADSC Q4 CCD detector as reported elsewhere (27Simmons C.R. Hao Q. Stipanuk M.H. Acta Crystallogr. Sect. F. 2005; 61: 1013-1016Crossref PubMed Scopus (9) Google Scholar). Single wavelength anomalous diffraction (SAD) data on cryo-cooled Se-Met CDO crystals were collected at the selenium K-edge (peak) at the Cornell High Energy Synchrotron Source F2 station using an ADSC Q210 CCD detector. The x-ray wavelength was set at 0.9790 Å based on a fluorescence scan. A total of 360 1° frames (180 + 180 by inverse beam geometry with 5° wedges) were recorded from one crystal. Diffraction data from frozen CDO-cysteine co-crystals were also collected at the Cornell High Energy Synchrotron Source F2 station. All data were reduced using the HKL package (29Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref Scopus (38432) Google Scholar); the data quality statistics are summarized in Table 1. All CDO crystals used in this study were isomorphous, belonging to space group P43212.TABLE 1Crystallographic data and refinement statistics NSLS, National Synchrotron Light Source; CHESS, Cornell High Energy Synchrotron Source; NA, not applicable; r.m.s.d., root mean square deviation.NativeSe-MetCys co-crystalData collection Experimental stationNSLS X12bCHESS F2CHESS F2 X-ray wavelength (Å)1.00.97900.9790 Exposure time (s)301520 Oscillation range (°)1.01.01.0 Cell dimensions (Å)a = b = 57.55, c = 123.06a = b = 57.49, c = 122.27a = b = 57.48, c = 122.80 Space groupP43212P43212P43212 Resolution (Å)1Values in parentheses are from the highest resolution shell30–1.50 (1.55-1.50)30–1.80 (1.86-1.80)30–1.50 (1.53-1.5) Unique reflections334531986532390 Multiplicity13.117.813.8 I/σ52.4 (7.0)31.8 (8.6)25.1 (1.4) Rsym (%)4.6 (31.8)8.7 (46.9)8.4 (71.3) Completeness (%)98.2 (95.8)100.0 (100.0)100.0 (100.0)Refinement R (%)/Rfree (%)18.0/20.8NA19.6/22.3 r.m.s.d. bonds (Å)0.02NA0.01 r.m.s.d. angles (°)2.06NA1.93a Values in parentheses are from the highest resolution shell Open table in a new tab Structure Determination and Refinement—The CDO structures were solved by SAD phasing. Four Se sites were located using the SAPI program (30Hao Q. Gu Y.X. Yao J.Y. Zheng C.D. Fan H.F. J. Appl. Crystallogr. 2003; 36: 1274-1276Crossref Scopus (9) Google Scholar). The correct space group, P43212, was selected by using the program ABS (31Hao Q. J. Appl. Crystallogr. 2004; 37: 498-499Crossref Scopus (66) Google Scholar) based on the four Se sites and SAD data to 3.0-Å resolution. The Se substructure was then fed into the program SOLVE (32Terwilliger T.C. Berendzen J. Acta Crystallogr. Sect. D. 1999; 55: 849-861Crossref PubMed Scopus (3219) Google Scholar) for refinement and phase calculation, resulting in an average figure of merit of 0.43 for all reflections between 20 and 2.3 Å. Combined with the native data set, the 2.3-Å SAD phases were gradually extended to 1.5 Å by solvent flipping implemented in the program SOLOMON (33Abrahams J.P. Leslie A.G. Acta Crystallogr. Sect. D. 1996; 52: 30-42Crossref PubMed Scopus (1141) Google Scholar), and an initial model accounting for 97% of the structure (r = 19.2 Rfree = 28.0) was automatically built with ArpWarp (34Perrakis A. Morris R. Lamzin V.S. Nat. Struct. Biol. 1999; 6: 458-463Crossref PubMed Scopus (2563) Google Scholar). Manual model building was performed in O (35Jones T.A. Zou J.-Y. Cowan S.W. Kjeldgaard M. Acta Crystallogr. Sect. A. 1991; 47: 110-119Crossref PubMed Scopus (13005) Google Scholar), alternating with crystallographic refinement using CNS (Crystallography and NMR System) software (36Brunger A.T. Adams P.D. Clore G.M. DeLano W.L. Gros P. Grosse-Kunstleve R.W. Jiang J.S. Kuszewski J. Nilges M. Pannu N.S. Read R.J. Rice L.M. Simonson T. Warren G.L. Acta Crystallogr. Sect. D. 1998; 54: 905-921Crossref PubMed Scopus (16944) Google Scholar) until final completion of the model. Rfree calculations were based on 10% of the reflections. During further refinement, water molecules were added in places with difference density >3.5 rrms, 2Fo - Fc density >1.0 rrms and having a reasonable environment. A close approach of Cys93-Sγ to Tyr157-Cϵ2 indicated the presence of a covalent link, and this led us to loosely restrain it to a bond length of 1.93 Å (based on an Sδ-Cϵ bond distance of methionine). Refinement was terminated when the remaining significant difference peaks were associated with alternate conformations of some water sites and of a few disordered side chains that were not near the active site (His20 and Val36). When refinement was complete, water molecules were numbered based on electron density strength, with Wat1 having the strongest density and Wat339 the weakest. The structure of the CDO-cysteine co-crystal was solved by difference Fourier, using as the initial model the final native CDO structure with active site waters removed. Clear movements indicated for the side chains of Arg60, Cys164, and Met179 and the backbone near Cys164 were accounted for manually, but solvent structure in the active site pocket and the density for a molecule apparently covalently attached to Cys164 were initially left uninterpreted. As refinement progressed, it became apparent that the active site was a mixture of a minor component indistinguishable from the native structure and a major component having Cys164 in an apparent disulfide link with an unknown ligand. Given this mixture, we decided to model the residual active site density (including the disulfide-linked sulfur site) as a series of water sites at the significant density maxima even though these sites were in some cases too close to each other and to protein atoms. These waters are numbered 401-428. Some residual density for the original native positions of Met179 and Arg60 remained. Final statistics for both refined models are given in Table 1. Coordinates—The atomic coordinates and structure factors of native CDO and the CDO-cysteine co-crystal have been deposited in the Protein Data Bank (PDB) with accession codes 2B5H and 2GH2, respectively. Overall Structure of CDO—The crystal structure of R. norvegicus CDO, solved by SAD phasing using Se-Met-substituted protein, yielded a final refined model with r = 18.0 and Rfree = 20.8 at 1.5-Å resolution. A total of 186 of the 200 residues in the protein (residues 5-190) were well defined in the electron density map (Fig. 1) and are included in the final model. No non-Gly residues have outlier ϕ,Ψ angles, and there is one cis-peptide preceding Pro159. The structure as a whole is highly similar to that of murine CDO (PDB accession code 2ATF; 100% sequence identity; root-mean-square deviation = 0.2 Å), including the cysteinyl-tyrosine linkage (Fig. 1). The only salient difference involves the metallocenter, as further discussed below. Briefly, the overall structure of CDO consists of a small α-helical domain containing three α-helices near the N terminus followed by 13 β-strands subdivided into a main β-sandwich domain and two β-hair-pins at the C terminus (Fig. 2). A short 310 helix is observed between β1 and β2. The entire β-sandwich is composed of seven anti-parallel β-strands (β1, β2, β4, β7, β9, β12, and β13) on the lower side and six anti-parallel β-strands (β3, β5, β6, β8, β10, and β11) on the upper side. N-terminal helices pack against the outside of the lower face of the sandwich to build a second non-polar core. Alignment of CDO sequences across multiple species reveals that the elements of secondary structure seen in the core β-sandwich of rat CDO are conserved in other CDOs, with all insertions and deletions occurring between the secondary structural elements (Fig. 3). In contrast, the C-terminal β-hairpins may be dispensable. The sequence alignment also reveals notably strong conservation of 18 residues (yellow in Fig. 3). Among these, 14 are present in the active site area, and their roles will be discussed below. The remaining four residues are likely important for structural reasons: Gly100 at the end of cupin motif 1 at a corner of the β-sandwich has a Gly-specific conformation of ϕ,Ψ = (+85, -157); Asn144, the terminal residue of cupin motif 2, is fully buried, stabilizing a loop spanning residues 144-150 by hydrogen bonding to three backbone groups (Ser146-N, Gly78-O, and Glu149-O) and a water molecule; Asn61 is fully buried and hydrogen bonds to Ser183-Oγ, Thr59-Oγ and the Ile74-O, apparently significant for formation of the zigzag chain path of segment 180-188 that contains two of the metal ligands; and Asn67, with ϕ,Ψ = (+55, +28), is in a G-N-G tripeptide with each residue in the α-L conformation, creating a short left-handed 310 helix. CDO Active Site—The CDO active site is identified by the mononuclear iron center that is fully occupied and is coordinated via a roughly tetrahedral geometry by the conserved residues His86, His88, His140 and the water molecule Wat4 (Figs. 1 and 4A). The B-factors for the iron (12 Å2) and Wat4 (10 Å2) are similar to those of the coordinating His residues (11-14 Å2), indicating full occupancy of the metal and no heterogeneity in the coordination geometry. The metallocenter is located in the central portion of the cupin β-sandwich (Fig. 2), consistent with what has been observed in other cupin structures. This metallocenter geometry appears different from the hexacoordinated Ni(II) center seen in the M. musculus CDO structure (26McCoy J.G. Bailey L.J. Bitto E. Bingman C.A. Aceti D.J. Fox B.G. Phillips Jr. G.N. Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 3084-3089Crossref PubMed Scopus (155) Google Scholar), but the difference can be seen simply to involve the additional presence in M. musculus CDO of two waters with long coordination distances (Fig. 4B). However, although geometrically small, this difference has major ramifications for mechanistic proposals (see “Discussion”). The iron is roughly 8 Å away from the protein surface and is surrounded by a solvent-filled pocket that connects to bulk solvent (Fig. 5). The remaining 14 conserved residues not mentioned above are all associated with the active site pocket, either lining it or adjacent to it (Fig. 5). Conspicuous among the conserved residues within the substrate binding pocket is Tyr157. Tyr157-OH forms a short (2.6 Å) hydrogen bond with Wat4, and as observed previously by McCoy et al. (26McCoy J.G. Bailey L.J. Bitto E. Bingman C.A. Aceti D.J. Fox B.G. Phillips Jr. G.N. Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 3084-3089Crossref PubMed Scopus (155) Google Scholar), very clear electron density shows that Cys93-Sγ is covalently bonded to Tyr157-Cϵ2 forming a cysteinyl-tyrosine linkage (Fig. 1). That Cys93-Sγ and Cys93-Cβ are roughly coplanar with the aromatic ring of Tyr157 indicates that the Cys93-Sγ to Tyr157-Cϵ2 bond has partial double bond character. This geometry was first observed by Ito et al. (37Ito N. Phillips S.E. Stevens C. Ogel Z.B. McPherson M.J. Keen J.N. Yadav K.D. Knowles P.F. Nature. 1991; 350: 87-90Crossref PubMed Scopus (686) Google Scholar) in galactose oxidase and more recently by Schnell et al. (38Schnell R. Sandalova T. Hellman U. Lindqvist Y. Schneider G. J. Biol. Chem. 2005; 280: 27319-27328Abstract Full Text Full Text PDF PubMed Scopus (86) Google Scholar) in NirA, a sulfite reductase. These are the only other structurally known examples of proteins containing a cysteinyl-tyrosine linkage. Additional highly conserved residues directly lining the active site include Tyr58, Arg60, Ser153, and His155 (Fig. 6). Tyr58 and Arg60 H-bond to waters in the active site and are thus well positioned to be directly involved in substrate coordination/catalysis. Ser153-Oγ H-bonds to His155-Nϵ2 (2.68 Å), and His155-Nδ2 H-bonds to Tyr157-OH (2.69 Å), forming a Ser153·His155·Tyr157 triad reminiscent of the Asp·His·Ser catalytic triad in chymotrypsin-like serine proteases (39Kraut J. Annu. Rev. Biochem. 1977; 46: 331-358Crossref PubMed Scopus (1073) Google Scholar, 40Dodson G. Wlodawer A. Trends Biochem. Sci. 1998; 23: 347-352Abstract Full Text Full Text PDF PubMed Scopus (529) Google Scholar). The conservation of Leu154, buried in a neighboring aliphatic pocket, and cis-Pro159-Pro160, located in a loop between β9 and β10, would seem related to ensuring accurate positioning of this triad of residues. Similarly, the conservation of Ser83, which H-bonds to the backbone NH of residue 142 (very close to the metal ligand His140), and Phe167, packed behind the main chain containing the metal ligands His86 and His88, may play a role in maintaining the integrity of the metallocenter. In addition, Asp87 is positioned between the iron-coordinating residues His86 and His88, and the Asp87 carboxylate interacts electrostatically with Asp87-N and Thr89-N, maintaining the position of the iron ligands, as well as with the backbone and side chain of His165 in the neighboring β-strand that contributes to the active site. Other residues that provide a non-polar lining to the active site pocket are Leu75, Trp77,Val142, Phe161, Cys164, Val177, and Met179. Overall, the pocket corresponds reasonably well to the space that would be required for a single molecule of cysteine. In addition to this main pocket, there is a smaller pocket located behind Wat4 immediately adjacent to Cys9